Training Requirements, Automation, and Job Polarisation
分析了工作培训要求如何与工程复杂性相互作用，影响企业的自动化决策，并用美国1980-2008年数据验证了培训要求和工程复杂性对就业极化的解释力。
Abstract We analyse how job training requirements interact with engineering complexity in shaping firms’ automation decisions. A model that distinguishes between a task’s engineering complexity and its training requirements predicts that when two tasks are equally complex, firms automate the task that requires more training. Under plausible conditions this leads to job polarisation, and in particular to polarisation of employment by initial training requirements. US data provide empirical support for the model’s implications. Training requirements and a measure of engineering complexity account for much of US job polarisation from 1980 to 2008.